🧠 Musk’s new “Scary Smart” AI

Also: Anthropic’s next AI model is almost here

Author

Evolving AI
February 17, 2025

Welcome, AI enthusiasts

Elon Musk said his startup xAI will release its Grok 3 chatbot on Monday and billed it as the "smartest AI on Earth" and “scary smart”. Let’s dive in! 

In today’s insights:

  • Musk’s “Scary Smart” AI comes Monday

  • Anthropic’s next AI model is almost here

  • Google’s Veo 2 now in YouTube Shorts

Read time: 4 minutes

LATEST DEVELOPMENTS

Source: Alamy

Evolving AI: Elon Musk says xAI’s Grok 3 chatbot will launch Monday with a live demo at 8 p.m. PT.

Key Points:

  • Musk says Grok 3 is the “smartest AI on Earth” and beats anything released so far.

  • The chatbot learns from its mistakes and fixes wrong information by itself.

  • xAI wants to hire thousands of people this year to make it even better.

Details:

Elon Musk says Grok 3 is coming Monday and will be shown live. He calls it “scary smart” and says it does better than any AI out there. “Grok 3 has very powerful reasoning capabilities, so in the tests that we’ve done thus far, Grok is outperforming anything that’s been released, that we’re aware of, so that’s a good sign,” he said. Musk first planned to wait a bit longer but decided to launch now after making final changes with his team. This comes after DeepSeek’s AI made big news in China for being as strong as US models but much cheaper. xAI started in 2023 and is growing fast with plans to add thousands of new workers. Grok 3 is trained at the Colossus Supercluster in Memphis with its 100,000 GPUs.

Why It Matters:

If Grok 3 is as smart as Musk says, it could change the AI race. But big claims bring big pressure — will it really be that good? If we remember last week, he put in an offer of $97B to buy OpenAI.

Source: Anthropic

Evolving AI: Anthropic is getting ready to launch a new AI model that can switch between deep thinking and fast answers.

Key Points:

  • The model can change between deep reasoning and quick replies, so developers can control costs.

  • Reports say it beats OpenAI’s o3-mini-high model in some coding tasks.

  • It works well with big codebases and business-related tests.

Details:

Anthropic’s new AI model is coming soon. It can switch between deep thinking and fast responses, making it more flexible. The model will also have a “sliding scale” so developers can decide how much computing power they need. Reports say it does better than OpenAI’s o3-mini-high model in some programming tasks and is great at working with large codebases. CEO Dario Amodei hinted at new models, saying he doesn’t see a clear line between normal AI and reasoning AI.

Why It Matters:

This a big move for Anthropic as the company has been quietly working in the backdrop as the Sam Altman led-OpenAI rolled out AI models one after another. Anthropic has a different approach — they’re pushing for responsible AI development. At the Paris AI Action Summit, CEO Dario Amodei warned about security risks, misuse by authoritarian regimes, and economic disruption. He stressed the need for strong leadership, supply chain security, and fair AI benefits. “Time is short, and we must act fast,” he said.

Source: YouTube

Evolving AI: YouTube Shorts now lets users make AI video clips with Veo 2, Google's latest model.

Key Points:

  • Dream Screen now works with Veo 2 to create better AI videos.

  • Users can make AI clips with text prompts and add them to.

  • Shorts Videos look more real with smoother movement and style options.

Details:

YouTube is upgrading Dream Screen with Veo 2, a new AI model from Google. Now users can type a prompt and get an AI-generated clip to use in Shorts. The update makes videos look more natural with better motion and style choices. Dream Screen also works faster than before. The feature is out now in the US, Canada, Australia, and New Zealand and more countries will get it later. The videos will have SynthID watermarks to show they are AI-generated.

Why It Matters:

Creators can now use AI to make full video clips, not just backgrounds, bringing new ways to tell stories on YouTube Shorts. This could change how people make content, but it also makes it harder to tell what is real and what is AI-generated.

 👀 Click on the image you think is real

QUICK HITS

🐶 Researchers are training AI to interpret animal emotions.

📵 Downloads of DeepSeek’s AI apps paused in South Korea over privacy concerns.

🧬 AI model deciphers the code in proteins that tells them where to go.

🏦 AI-generated content raises risks of more bank runs, UK study shows.

🍏 Apple aims to bring AI features and spatial content app to Vision Pro.

🔍 Perplexity launches its own freemium ‘deep research’ product.

🏥 xAI’s “Colossus” supercomputer raises health questions in Memphis.

🇨🇳 China’s Tianjin city adopts DeepSeek as part of rush to embrace domestic AI industry.

📈 Trending AI Tools

  • 🔍 Legal Graph - A tool to visualize connections between legal concepts, cases, statutes, and regulations (link)

  • 🤖 SWE-Agent - A tool to autonomously resolve bugs in GitHub repositories (link)

  • 📧 Emilio - A tool to organizes and prioritizes email inbox, summarizes threads, and drafts responses (link)

  • 📝 PowerDreamer - High-quality writing assistance across a wide range of needs, making it an essential tool for professionals, content creators, and job seekers alike (link)

  • 📊 PrometAI - Revolutionize business planning with AI, strategic tools, and precision valuation (link)

Reply

or to participate.